Inflation and employment -- this is what the contest is actually judging you on. Everything else in this briefing is supporting context.
Computed divergences in the current data -- observations, not conclusions. What they mean, and whether they belong in your analysis, is your call.
Consider: Worth digging into whether the gap is concentrated in a few sticky categories (shelter, services) or broad-based -- that distinction usually drives how a real FOMC statement gets worded.
Consider: A gap this size usually means either the market thinks the current print is noisy/transitory, or forward guidance hasn't caught up to the data yet -- worth having a view on which.
Consider: Worth checking whether this lines up with a known pressure point (month-end, tax date, Treasury settlement) or looks like broader funding stress -- the difference matters for whether it's noise or signal.
One worked example, built the way past winners built theirs -- not a template to copy, a benchmark to compare against.
Two independently-built forecasts landed within a few tenths of each other and of the FOMC's own median -- the real disagreement below is about the policy response to a shared economic outlook, not about the outlook itself.
| Variable | Period | FOMC SEP median | Our forecast |
|---|---|---|---|
| Real GDP growth | 2026 | 2.3% | 2.4% |
| Private domestic final purchases (+4.2% annualized, Q2) and a capex boom (core capital-goods orders +12.6% YoY) are running hotter than the SEP assumes. | |||
| Real GDP growth | 2027 | 2.4% | 2.0% |
| Slower as the energy shock and October tightening work through; consistent with the SEP's own path decelerating from '28 onward. | |||
| Unemployment rate | 2026 | 4.1% | 4.1% |
| Agree with the SEP -- August's rebound (+162k, participation firm) confirms full employment by every model estimate (KC Fed u* also 4.1%). | |||
| Unemployment rate | 2027 | 4.1% | 4.3% |
| A modest cost from the extra tightening step -- the SEP holds flat throughout, we think that's slightly optimistic given the October move. | |||
| Core PCE inflation | 2026 | 3.4% | 3.5% |
| Above the SEP -- three-month annualized core PCE is already running at 3.05% with no deceleration in the Cleveland Fed nowcast. | |||
| Core PCE inflation | 2027 | 2.5% | 2.6% |
| Converges close to the SEP as the October step and energy-shock fade both work through by year-end 2027. | |||
| Fed funds rate (year-end) | 2026 | 4.10% | 4.00–4.25% |
| We agree with the Committee's own median that one more step is appropriate this year -- and think it should come at the next meeting, not be deferred. | |||
| Fed funds rate (year-end) | 2027 | 4.10% | 4.00–4.25% |
| Hold after the October step, consistent with the SEP's own median path for 2027. | |||
Dated thresholds, not a moving target -- if a line is crossed before the next meeting, revisit.
| Indicator | Review line | Latest | Read |
|---|---|---|---|
| Core PCE, 3-month annualized | hold line 3.0% | 3.05% | Above -- still triggers action |
| Cleveland Fed 10y expected inflation | hold line 2.5% | 2.57% | Above -- anchor still drifting |
| PPI final demand, 3-month annualized | further-hike review 4% | 1.4% | Within line |
| Initial claims, 4-week average | easing review 240k | 203k | Within line |
| High-yield OAS | easing review 3.20pp | 2.68pp | Within line |
| Nonfarm payrolls, 3-month average | easing review below zero | +71k | Within line |
Map: our forecast against the Fed's own, then the case for acting in October, then a real disagreement about what that case is missing, then how we resolve it, then the vote. On financial conditions: the ten-year real (TIPS) yield is 2.68%, the highest since 2008, yet the Chicago Fed's own financial conditions index sits at -0.56 and high-yield spreads are historically tight at 2.68 points -- bonds have tightened, risk assets haven't. That gap matters for October: policy transmission through risk appetite hasn't happened yet, which is itself part of the case for finishing the move now rather than assuming past tightening will keep doing the work.
Core PCE -- the measure the Fed's 2% objective is actually defined on -- is running at 3.05% three-month annualized, above the 3.0% threshold a disciplined team should set as its own hold line, with no deceleration in the Cleveland Fed's nowcast ahead of the Sept 30 release. Long-horizon expectations are drifting: Cleveland Fed's 10-year expected inflation is 2.57%, the highest since 2007. Per BIS research, the core-inflation effect of an energy shock more than doubles when expectations sit above target, and the real policy rate is one of three variables that condition that pass-through -- a neutral real rate (1.23%, squarely inside the Fed's own 1.0-1.7% neutral band) during an active supply shock is the exposed position, not the cautious one. And waiting for confirmation doesn't actually buy safety: Dallas Fed high-frequency evidence shows spending responds to a policy shock within weeks, while consumer prices don't trough for eleven months -- meaning October data will show us the demand response to holding, not the price response, well before we'd have real confirmation either way. With the labor market at full employment by every yardstick (unemployment 4.1% matches the Kansas City Fed's own u* estimate exactly) and low turnover removing the wage-spiral channel, a 25bp step costs little in employment terms. This isn't a call to keep hiking indefinitely -- it's a call to finish one step now, while it's cheap.
Presenter 2's own numbers cut the other way for a specific group, and I don't think we get to call a hike 'cheap' without accounting for that. Real hourly earnings have been negative for five straight months. The lowest wage quartile is growing at 3.7%, no longer meaningfully ahead of 3.4% inflation -- the buffer that used to protect the bottom of the distribution is gone. The bottom 50% of households hold just 2.3% of net worth, and the Fed's own SHED survey finds 40% of adults under $50,000 couldn't cover a $100 emergency from savings. For them, price stability isn't an abstraction, but neither is a labor market that gets even marginally tighter to enter. Teen unemployment is 14.1%; Black unemployment is 6.0% -- both far above the 4.1% headline Presenter 2 is calling 'full employment.' This is a low-hire, low-fire equilibrium: hiring is running below its 2015-19 average, which means when a policy mistake does show up, it shows up first among teenagers, new graduates, and Black workers -- the people a low-hire market reaches last and fails first. And the housing channel is already doing real work without our help: 30-year mortgage rates at 6.81% have largely locked out first-time buyers already. Tightening further doesn't just risk being unnecessary -- it risks being distributionally regressive for a gain that's still uncertain, since the 5-year-5-year forward breakeven -- the market's own longer-run inflation read -- is still anchored at 2.33%, 'still ordinary' by historical standards.
This is the one section where I'd point the team at this dashboard's own Second District panel rather than a fixed number -- Empire State's survey updates monthly, and whatever it shows as of your actual presentation date is more current than anything written here. Structurally: regional manufacturing data is the kind of evidence that either independently corroborates the national story or complicates it -- check it before assuming it settles the disagreement between Presenter 2 and Presenter 3 either way.
Presenter 2 and Presenter 3 are both arguing from the same dataset and neither is wrong -- this is a genuine weighting question, not a data question, and it deserves an honest resolution rather than picking a side and ignoring the other's evidence. We side with Presenter 2 on the immediate decision: the Fed's own target gauge has already crossed a disciplined hold line, and the Dallas Fed timing evidence means waiting doesn't actually de-risk the call the way it feels like it should. But Presenter 3's evidence changes what we commit to alongside the hike, not just how we feel about it. We're proposing the Committee adopt explicit, quantified distributional review lines -- teen, Black, and new-graduate unemployment; real wage growth in the lowest quartile -- with the same seriousness as the inflation and growth lines already in the SEP, and a standing commitment to pause further tightening if they cross defined thresholds. That's the actual synthesis: act on the anchoring risk now, while making sure the cost of being wrong is something we're accountable for tracking, not something that just quietly shows up first in the groups least able to speak up about it.
FRED, most recent observation per series.
| Indicator | Latest | Value | Change |
|---|---|---|---|
| CPI (All Urban Consumers) | 2026-08-01 | 334.13 | ▲ 1.32 |
| Core PCE Price Index | 2026-08-01 | 130.46 | ▲ 0.32 |
| Unemployment Rate | 2026-08-01 | 4.10 | — 0.00 |
| Nonfarm Payrolls | 2026-08-01 | 159,075.00 | ▲ 162.00 |
| Real GDP | 2026-04-01 | 24,408.01 | ▲ 133.63 |
| Fed Funds Effective Rate | 2026-08-01 | 3.63 | — 0.00 |
| 2-Year Treasury Yield | 2026-09-29 | 4.89 | ▼ 0.03 |
| 10-Year Treasury Yield | 2026-09-29 | 5.26 | ▲ 0.02 |
| 20-Year Treasury Yield | 2026-09-29 | 5.64 | ▲ 0.04 |
| 30-Year Treasury Yield | 2026-09-29 | 5.59 | ▲ 0.03 |
| Initial Jobless Claims | 2026-09-19 | 197,000.00 | ▼ 1,000.00 |
| University of Michigan Consumer Sentiment | 2026-08-01 | 51.70 | ▼ 3.50 |
| Empire State Mfg Survey (General Business Conditions) | 2026-09-01 | 7.60 | ▼ 13.00 |
The plumbing behind QT/QE, reserve management purchases, and Standing Repo Facility decisions -- SRF usage should sit near zero on a normal day; a jump here is real, current funding stress.
| Indicator | Latest | Value | Change |
|---|---|---|---|
| Effective Fed Funds Rate (daily) | 2026-09-29 | 3.88 | — 0.00 |
| Interest on Reserve Balances (IORB) | 2026-10-01 | 3.90 | — 0.00 |
| SOFR | 2026-09-29 | 3.88 | ▼ 0.02 |
| SRF Usage | 2026-09-30 | $1.2B | $1.2B |
| ON Reverse Repo Usage | 2026-09-30 | $11.5B | $0.1B |
Empire State Manufacturing Survey (seasonally adjusted diffusion indexes) -- your differentiator against teams running purely national analysis.
| Indicator | Latest | Value | Change |
|---|---|---|---|
| General Business Conditions | 2026-09-30 | 7.6 | ▼ 13.00 |
| New Orders | 2026-09-30 | 2.0 | ▼ 15.30 |
| Shipments | 2026-09-30 | -3.2 | ▼ 14.90 |
| Prices Paid | 2026-09-30 | 63.1 | ▲ 4.50 |
| Prices Received | 2026-09-30 | 28.1 | ▲ 5.40 |
| Number of Employees | 2026-09-30 | 10.6 | ▲ 1.30 |
yfinance, ~15-minute delayed.
| Indicator | Symbol | Last | Prior | % Chg |
|---|---|---|---|---|
| S&P 500 | ^GSPC | 7,651.54 | 7,670.84 | ▼ 0.25% |
| 10Y Treasury Yield (index) | ^TNX | 5.29 | 5.25 | ▲ 0.72% |
| 30Y Treasury Yield (index) | ^TYX | 5.64 | 5.59 | ▲ 0.79% |
| US Dollar Index | DX-Y.NYB | 101.50 | 101.37 | ▲ 0.13% |
| Crude Oil (WTI) | CL=F | 89.98 | 89.38 | ▲ 0.67% |
| Gold | GC=F | 4,186.40 | 4,179.70 | ▲ 0.16% |
No FOMC statement supplied -- rerun with --fomc-statement to include this section.
Public record from federalreserve.gov, paraphrased in our own words -- for studying approach, not for copying into your script.
Framed the entire presentation around one tension: 'an unbalanced dual mandate' where downside labor-market risk outweighed upside inflation risk from new tariffs. Walked output, labor, inflation, and financial conditions, then chose between two explicitly named policy options -- hold, or cut 25 basis points -- landing on the cut.
Explicitly updating their own recommendation live, with the specific new data that changed it, is about as strong a demonstration of 'data-dependent' thinking as is possible in a Q&A -- it shows the analysis was real, not just memorized and defended no matter what.
Fluently distinguished short-run vs. long-run r-star on a cold follow-up question with no slide to lean on -- exactly the 'thorough understanding of sophisticated concepts' the top rubric tier calls for, tested live rather than just rehearsed.
Constant explicit signposting ('conclusively... contributing to the downside risks... rounding out the picture...') kept a dense 15 minutes of data organized and easy for judges to follow along in real time.
Framed the whole 15 minutes as a single live FOMC-style debate about how fast to keep cutting rates after starting an easing cycle. Walked through growth, then labor, then inflation, then financial conditions, in that order, before landing on a specific vote.
Their evidence wasn't just data points -- it was data points attributed to specific, real, current voices (a named Fed president's public comments, a named academic's recent research). That's precisely what the rubric's top tier asks for: 'a wide variety of authoritative sources,' not just charts.
Structuring genuine on-mic disagreement between teammates is a hard way to present, but it's the clearest possible evidence against the rubric's warning sign of 'one or two team members dominate.' Every presenter both made a claim and pushed back on a teammate's claim at least once.
The explicit agenda at the very start (what topics, in what order, ending in a vote) gave judges a map before the content even began -- exactly the 'logical and coherent organization' the top rubric tier calls for.
Not directly observable from the presentation alone, but the format itself -- team members trained to argue live against their own teammates' positions -- is a natural way to build the quick, poised rebuttal skill the Q&A round specifically scores.
Recommended holding the fed funds rate at 1.5-1.75% amid slowing global growth and below-target inflation, but didn't stop at a simple hold -- proposed two specific new policy tools: a form of temporary price-level targeting with a one-year lookback, and a standing repo facility to fix reserve-distribution problems that had caused a rate spike months earlier.
Proposing two specific, named policy instruments rather than a plain hold/cut/hike decision shows original synthesis of the research, not just reporting it -- a step beyond what most teams likely offered.
Surfacing your own recommendation's weakness before being asked, then explaining the specific design choice made to address it, is a strong signal of genuine understanding rather than a rehearsed pitch -- much harder to fake under follow-up.
Multiple team members visibly building on one answer across a multi-turn follow-up, rather than one person carrying the exchange, is exactly the 'substantial and integral role' for every member the rubric rewards.
Both teams recommended holding rates. Pace went further and proposed two specific, named new tools (a price-level-targeting variant, a standing repo facility) to address weaknesses they identified in the current framework. Harvard's structural proposal (organic balance-sheet growth via 1-2 year Treasuries) addressed one issue rather than two -- a narrower scope of original policy design in the same time slot.
Harvard's answer was theoretically solid -- wage growth reaching lower-income workers late in the cycle, a wealth effect concentrated among asset holders. Pace made a similar theoretical case but added a specific, named real-world anchor (a sitting Fed Bank president's own account of a 'Fed Listens' community event) -- grounding the same argument in a concrete, current example rather than theory alone.
Both teams were pushed hard on the wage-inflation (Phillips curve) relationship. Pace's exchange involved three different team members adding distinct pieces to the answer across the follow-up chain; Harvard's corresponding exchange was comparatively more concentrated in fewer voices. Neither answer was wrong -- but Pace's version demonstrated shared command of the topic across more of the team.
Six patterns that show up across every case study above, not just once -- each one traceable back to a specific transcript, not generic advice.
Every single case study did this. 2024 Princeton tied claims to a named Fed president's public remarks and named academic papers. 2025 Pace cited specific FOMC governors' recent speeches by name. 2019 Pace and Harvard both built arguments around named authors' research. A statistic on a slide is forgettable; 'as President Logan noted last month' is what the rubric means by 'authoritative sources' and what makes an answer defensible under follow-up.
2019 Pace didn't just recommend holding rates -- they designed two named new tools (a price-level-targeting variant, a standing repo facility). 2025 Pace revised a specific balance-sheet/QT recommendation, not just the headline rate. A plain hold/cut/hike is table stakes; engaging with the how -- QT pace, new facilities, forward guidance design -- is what separated winners from a merely correct rate call.
2024 Princeton's team argued with each other on mic. 2019 Pace had three different members build on one answer across a multi-turn follow-up, versus a narrower set of voices in Harvard's equivalent exchange that same day. The common thread: winning teams distribute both the content and the friction across the whole team, instead of one strong presenter carrying everyone else.
2025 Pace openly changed a recommendation mid-Q&A when asked what data would change their mind, and explained exactly why. 2019 Pace named their own recommendation's weakness before a judge could catch it. Defending an original slide no matter what reads as rehearsed; visibly updating your view under new information reads as genuine understanding -- which is what this rubric category is actually trying to measure.
The clearest single difference in the 2019 head-to-head: both teams gave theoretically sound answers on inequality, but Pace anchored theirs in a specific, named Fed president's own account of a real community event. Same argument, but one version was textbook and one was concrete -- concrete reads as deeper understanding even when the underlying theory is identical.
2024 Princeton opened with an explicit agenda before any data appeared. 2025 Pace narrated its own structure throughout ('conclusively... rounding out the picture...'). Judges are tracking five different presentations in a row -- narrating your own organization, not just having good organization, is what keeps a dense 15 minutes legible to someone hearing it once.